2016 Volume 4 Issue 2 Pages 88-96
Aim: To evaluate whether there are any women with postpartum hemorrhage (PPH) showing a hemoglobin (Hb) concentration of <7.0 g/dl in cases with estimated blood loss of <1,500 ml or a shock index (SI) of >1.5 in cases with estimated blood loss of <2,500 ml.
Methods: We reviewed the records of 36 women transferred due to PPH in our tertiary center in 2002-2005. We collected the patients’ information including the estimated blood loss, Hb concentration, and SI on arrival.
Results: In the group of 500–1,499 ml, 13% (2/16) showed Hb concentrations of <7.0 g/dl. In the group of 1,500–2,499 ml, one woman with uterine inversion whose estimated blood loss was 1,600 ml showed SI of 1.9 with a Hb concentration of 6.3 g/dl. SI in women with uterine inversion was significantly higher than in those with atonic bleeding, retention of placenta/accreta, and vaginal/cervical lacerations, respectively. SI in all women with uterine inversion was >1.0, although in three quarters of the cases, the blood loss was reported as <2,500 ml.
Conclusions: Our chart review clearly demonstrates that there are some PPH cases showing mismatch between the reported estimated blood loss and the Hb concentration and/or SI. Especially, SI may be clinically useful for judging transfusion in women with uterine inversion. We should identify the factors in such mismatching cases, and reveal the pitfalls of the current guideline for critical obstetrical hemorrhage mainly based on SI.
Postpartum hemorrhage (PPH) is a leading cause of maternal mortality in many countries, including Japan.1,2) In a detailed assessment of the preventability of 197 in-hospital maternal deaths in Japan in 1991–1992, 70% or more of obstetrical specialists judged that 72 cases (37%) should not have been difficult to prevent.2) Of these 72 preventable cases, 64% were due to ante- and PPH.2) Since PPH is often preventable and treatable, the “Three Delays” model of maternal mortality has been proposed as the cause of delayed treatment for PPH: a delay in the decision to seek care, a delay in reaching care, and a delay in receiving quality care.3) All forms of delay sometimes occurred in cases with obstetrical hemorrhage, because the estimated blood loss was very inaccurate in emergency cases with marked rapid bleeding after delivery.
In Japan, the guideline for critical obstetrical hemorrhage was published in 2010 to promote standard blood transfusion methods for pregnant and puerperal women with excess blood loss.4,5) In the guideline, a shock index (SI), but not estimated blood loss, is introduced to make a decision regarding clinical performance and the start of transfusion.4,5) Estimated blood loss may be underestimated rather than actual blood loss, because not all blood can be collected into the blood-collecting bag on the delivery bed, and because it is very difficult to measure blood loss in an ambulance during the transportation of patients from primary clinics/hospitals to an emergency hospital.6,7) On the contrary, in some cases, estimated blood loss may be overestimated rather than actual blood loss, because amniotic fluid easily contaminates the blood in cases of cesarean section,8) and because the massive infusion of crystalloid/colloid solutions for women with PPH theoretically dilutes the blood.9) The latter is especially important because fluid is usually replaced at the primary facility in almost every transferred patient due to PPH in Japan.
Are there any cases in which estimated blood loss before arrival at a tertiary center is underestimated rather than actual blood loss? The averaged body weight of Japanese women aged 20–24 years old is almost 50 kg. As the blood volume in non-pregnant women is 78 ml/kg, a blood volume in a woman weighing 50 kg is estimated to be 3,900 ml. Since the blood volume during pregnancy increases on average by 25%, the blood volume in the third trimester is estimated to be 4,875 ml. Let blood loss of 1,500 ml occur after delivery. Because the mean hemoglobin (Hb) concentration in the third trimester is almost 11.0 g/dl,10,11) the Hb concentration after bleeding of 1,500 ml in average Japanese pregnant women is estimated to be 7.6 g/dl under the appropriate infusion of crystalloid and/or colloidal solutions (calculated by the equation of 11.0×[4,875–1,500]/4,875). Then, we hypothesized that if estimated blood loss is accurately assessed, the Hb concentration on arrival at a tertiary center will be ≥7.0 g/dl in all women with PPH, in whom the estimated blood loss is reported as <1,500 ml. A Hb level of 7 g/dl is the most commonly used threshold for red blood cells (RBC) transfusion in the clinical practice guidelines for RBC transfusion.12,13) In the treatment guideline for critical obstetrical hemorrhage, it is stated that for pregnant women, SI of 1.0 is estimated as blood loss of 1,500 ml, and SI of 1.5 is estimated as blood loss of 2,500 ml.5) Thus, we also hypothesized that if blood loss is accurately estimated, SI on arrival at a tertiary center will be <1.5 in all women with PPH, in whom the estimated blood loss has been reported as <2,500 ml. To assess the two above-mentioned hypotheses, we reviewed our medical charts for women with PPH who were transferred from primary clinics/hospitals to our tertiary center in 2002–2005. We extracted information including the blood loss, Hb concentration, and SI in 36 transferred women with PPH, and evaluated whether there were some cases showing Hb concentrations of <7.0 g/dl among those with estimated blood loss of 500–1,499 ml or SI of >1.5 in cases with estimated blood loss of 500–2,499 ml. Next, we classified the main causes of PPH into 4 groups: atonic bleeding, placental retention/accreta, vaginal/cervical lacerations, and uterine inversion, and evaluated whether the causes of PPH affect the vital signs, SI, and Hb concentrations on arrival at our tertiary center.
We collected the old data in 2002–2005, when SI was not being used for the management of obstetrical hemorrhage in our tertiary center. We reviewed 61 women with obstetrical hemorrhage after delivery who were transferred from primary clinics/hospitals to our tertiary center, and collected the patients’ information including estimated blood loss, Hb concentration, and vital signs (heart rate [HR], systolic blood pressure [SBP], and SI [HR divided by SBP]), which were measured within 10 min following the arrival of women with PPH. Values measured at the time of the highest SI were selected for analysis. The Hb concentration on the first blood sampling just after the arrival of the patients at our tertiary center was used. We also collected the following data: estimated blood loss before arrival, total blood loss, platelet counts on arrival and at the nadir, international normalized ratio of prothrombin time (PT-INR) on arrival and its maximum level, fibrinogen, fibrin degradation products (FDP) on arrival and its maximum level, RBC or fresh frozen plasma (FFP) transfusion, initial treatments before arrival, causes of PPH, presence/absence of hysterectomy, and maternal and infantile outcomes. After excluding women with abdominal/vulvar hematoma (n=17), those with estimated blood loss before arrival <500 ml (n=6), a woman with placental abruption on transvaginal delivery (n=1), and a woman with severe atonic bleeding who was adequately transfused at the primary clinic and then transferred to our intensive care unit (n=1), the remaining 36 women underwent analysis; all 36 women had PPH, and the main causes of PPH were atonic bleeding (n=14), placental retention/accreta (n=12), vaginal/cervical lacerations (n=6), and uterine inversion (n=4).
We also collected 28 women with obstetrical hemorrhage after delivery showing estimated blood loss ≥500 ml who were transferred from primary clinics/hospitals to our tertiary center in 2013–2015; they consisted of 14 women with atonic bleeding, 5 women with placental retention/accreta, 2 women with vaginal/cervical laceration, 3 women with uterine inversion, and 5 women with late PPH. These cases were used to examine whether obstetrical hemorrhage cases showing mismatch between the reported estimated blood loss and the Hb concentration and/or vital signs were decreased after sufficient recognition of the guideline for the management of critical obstetrical hemorrhage.
In pregnant women, SI of 1.0 is estimated to represent blood loss of 1,500 ml, and that of 1.5 is estimated to represent blood loss of 2,500 ml. In 2002–2005, our institutional indications for rapid RBC/FFP transfusions for women with PPH were as follows: (1) Hb concentration of <7.0 g/dl, (2) SBP of <80 mmHg, or (3) a complication of disseminated intravascular coagulation (DIC) according to the obstetrical DIC score;14,15) however, we did not use SI for the decision regarding transfusion in 2002–2005. We obtained approval from the Ethics Committee of our institute for this retrospective study in 2016.
StatisticsThe results are presented as the mean±SD. The continuous variables of multiple groups were analyzed using analysis of variance (ANOVA) followed by Bonferroni’s multiple comparisons. The categorical data of multiple groups were analyzed using the chi-square test followed by multiple comparisons with Bonferroni’s adjustment. All analyses were performed using the IBM SPSS software package (version 21, IBM, Armonk, NY, USA). A P-value <0.05 was considered significant.
The rates of women with estimated blood loss of 500–1,499, 1,500–2,499, and ≥2,500 ml were 44% (16/36 cases), 50% (18/36 cases), and 6% (2/36 cases), respectively. Among the three groups, the age, birth weeks, and birth weight were not different (Table 1). The frequencies of nulliparous women and cesarean section were also not different among them. As for the main cause of PPH, atonic bleeding was the most frequent in the group of 1,500–2,499 ml, whereas placental retention/accreta was the most common in the group of 500–1,500 ml. In most cases, the amounts of infusion were less than twice as much as the estimated blood loss (75% in 500–1,499 ml, 100% in 1,500–2,499 ml, and ≥2,500 ml, respectively; 89% in all women). RBC transfusion before arrival at our tertiary center was performed in only two cases; in a case with estimated blood loss of 1,595 ml, two units of RBC were transfused before arrival, and no other RBC was transfused after arrival; in another case with estimated blood loss of 2,750 ml, one unit of RBC was transfused before arrival, and 16 units of RBC, 15 units of FFP and 10 units of platelet concentrate were transfused after arrival. In 2002–2005, intrauterine tamponade or uterine compression suture was rarely used; instead, uterine/cervical gauze packing was often used.
| Estimated blood loss before arrival (ml) | Missing data | P-value | Significant pair(s) | ||||
|---|---|---|---|---|---|---|---|
| 500–1,499 | 1,500–2,499 | ≥2,500 | |||||
| (Group 1, n=16) | (Group 2, n=18) | (Group 3, n=2) | |||||
| Age, yr | 28.5±5.8 | 30.1±4.7 | 27.0±2.1 | 0 | 0.551 | ||
| Nulliparous women (%) | 13 (81) | 11 (61) | 1 (50) | 0 | 0.336 | ||
| Cesarean section (%) | 1 (6) | 3 (17) | 0 (0) | 0 | 0.688 | ||
| Weeks of delivery | 39.2±1.8 | 39.7±1.4 | 38.5±1.4 | 0 | 0.441 | ||
| Newborn weight (g) | 2,892±405 | 3,131±444 | 2,959±366 | 1 (group 1), 1 (group 2) | 0.290 | ||
| Main cause of PPH | |||||||
| Atonic bleeding | 1 (6) | 13 (72) | 0 (0) | 0 | 0.001 | 1 vs. 2 | |
| Placental retention/accreta | 7 (44) | 4 (22) | 1 (50) | ||||
| Vaginal/cervical lacerations | 6 (38) | 0 (0) | 0 (0) | 1 vs. 2 | |||
| Uterine inversion | 2 (13) | 1 (6) | 1 (50) | ||||
| Initial treatments before arrival | |||||||
| Fluid resuscitation ≥1,500 ml | 7 (44) | 14 (78) | 2 (100) | 0 | 0.066 | ||
| Fluid resuscitation ≥twice as much as estimated blood loss | 4 (25) | 0 (0) | 0 (0) | 0 | 0.060 | ||
| RBC transfusion | 0 (0) | 1 (6) | 1 (50) | 0 | 0.014 | 1 vs. 3 | |
| Uterine/Cervical gauze packing | 5 (31) | 5 (28) | 1 (50) | 0 | 0.808 | ||
| Holding the cervix | 1 (6) | 2 (11) | 0 (0) | 0 | 0.797 | ||
| Intrauterine balloon tamponade | 1 (6) | 0 (0) | 0 (0) | 0 | 0.526 | ||
| Uterine compression suture | 0 (0) | 0 (0) | 0 (0) | 0 | 1.000 | ||
Data are presented as mean±SD, or number (%).
In the group of ≥2,500 ml, HR and SI were significantly higher than in the group of 500–1,499 ml and that of 1,500–2,499 ml, respectively (Table 2). Hb concentrations on arrival at our tertiary center were significantly decreased in the group order of 500–1,490 ml >1,500–2,499 ml >≥2,500 ml. In the group of 500–1,499 ml in 2002–2005, 13% (2/16) of women showed a Hb concentration of <7.0 g/dl, whereas, in the group of 500–1,499 ml in 2013–2015, 17% (2/12) of women showed a Hb concentration of <7.0 g/dl. Especially, one woman in 2002–2005 showed a Hb concentration of 5.5 g/dl and her SI was increased to 1.38 (see arrow in Figure 1, left panel and Figure 2, left panel), and 4 units of RBC were finally transfused; she was complicated by uterine atony. In addition, one women in 2013–2015 showed SI of 1.7, but her Hb concentration was 8.4 g/dl (see solid circle in Figure 1, right panel and Figure 2, right panel), and RBC/FFP transfusion (total of 30 units and 26 units, respectively) was started soon after arrival; she was complicated by uterine inversion. In the group of 1,500–2,499 ml in 2002–2005, 22% (4/18) of women showed a Hb concentration of <7.0 g/dl, whereas, in the group of 1,500–2,499 ml in 2013–2015, 50% (5/10) of women showed a Hb concentration of <7.0 g/dl. Especially, one woman with an estimated blood loss of 1,600 ml in 2002–2005 showed SI of 1.9 with a Hb concentration of 6.3 g/dl (see solid circle in Figure 1, left panel and Figure 2, left panel), and RBC/FFP transfusion (total of 10 units and 10 units, respectively) was started soon after arrival; she was also complicated by uterine inversion. Taken together, the obstetrical hemorrhage cases showing mismatch between the reported estimated blood loss and the Hb concentration and/or SI still existed in 2013–2015.
| Estimated blood loss before the arrival (ml) | Missing data | P-value | Significant pair(s) | |||
|---|---|---|---|---|---|---|
| 500–1,499 | 1,500–2,499 | ≥2,500 | ||||
| (Group 1, n=16) | (Group 2, n=18) | (Group 3, n=2) | ||||
| HR, (per min) | 95±26 | 99±27 | 152±17 | 0 | 0.024 | 1 vs. 3, 2 vs. 3 |
| SBP (mmHg) | 119±25 | 115±22 | 75±21 | 0 | 0.054 | |
| SI | 0.83±0.28 | 0.88±0.29 | 2.11±0.36 | 0 | <0.001 | 1 vs. 3, 2 vs. 3 |
| SI ≥1.5 (%) | 0 (0) | 1 (6) | 2 (100) | 0 | <0.001 | 1 vs. 3, 2 vs. 3 |
| Hb on arrival (g/dl) | 9.8±2.0 | 8.0±1.4 | 3.1±2.2 | 0 | <0.001 | All pairs |
| Hb on arrival <7.0 (%) | 2 (13) | 4 (22) | 2 (100) | 0 | 0.019 | 1 vs. 3 |
| Plt on arrival <120×103/μL (%) | 0/12 (0) | 0/13 (0) | 0/1 (0) | 10 | – | – |
| PT-INR on arrival ≥1.25 (%) | 0/12 (0) | 1/12 (8) | 1/1 (100) | 10 | – | – |
| Fibrinogen on arrival <150 mg/dl (%) | 1/4 (25) | 1/6 (17) | 1/1 (100) | 25 | – | – |
| FDP on arrival ≥40 μg/ml (%) | 0/5 (0) | 1/10 (10) | 0/1 (0) | 20 | – | – |
| RBC transfusion (%) | 4 (25) | 7 (39) | 2 (100) | 0 | 0.127 | |
| FFP transfusion (%) | 2 (13) | 5 (28) | 2 (100) | 0 | 0.025 | 1 vs. 3 |
For Plt, PT-INR, fibrinogen, and FDP, we could not collect data before 2,004, resulting in many missing values. Therefore, we did not compare the data among the three groups.
Data are presented as mean±SD, or number (%).
HR, heart rate; SBP, systolic blood pressure; Plt, platelet; PT-INR, international normalized ratio of prothrombin time; FDP, fibrinogen degradation products; FFP, fresh frozen plasma.
Relationship between the estimated blood loss before arrival at our tertiary center and Hb concentration on arrival in 36 women with postpartum hemorrhage (PPH) in 2002–2005 (left panel) and 28 women with PPH in 2013–2015 (right panel).
The horizontal line indicates Hb of 7.0 g/dl. The left and right vertical lines indicate blood loss of 1,500 and 2,500 ml, respectively. The closed circle in the left panel indicates a woman with uterine inversion, who showed a Hb concentration of 6.3 g/dl and shock index (SI) of 1.90; the closed circle in the right panel indicates a woman with uterine inversion, who showed a Hb concentration of 8.4 g/dl and SI of 1.70; and the arrow in the right panel indicates a woman with atonic bleeding, who showed a Hb concentration of 5.5 g/dl and SI of 1.38. In 2002–2005, the fractions of women with PPH showing a Hb concentration of <7.0 g/dl with estimated blood loss of 500–1,499, 1,500–2,499, and ≥2,500 ml were 13, 22, and 100%, respectively. In 2013–2015, those fractions were 17, 50, and 67%, respectively.
Relationship between the estimated blood loss before arrival at our tertiary center and the shock index (SI) on arrival in 36 women with PPH in 2002–2005 (left panel) and 28 women with postpartum hemorrhage (PPH) in 2013–2015 (right panel).
The horizontal line indicates SI of 1.5. The left and right vertical lines indicate blood loss of 1,500 and 2,500 ml, respectively. The closed circle in the left panel indicates a woman with uterine inversion, who showed a Hb concentration of 6.3 g/dl and SI of 1.90; the closed circle in the right panel indicates a woman with uterine inversion, who showed a Hb concentration of 8.4 g/dl and SI of 1.70; and the arrow in the right panel indicates a woman with atonic bleeding, who showed a Hb concentration of 5.5 g/dl and SI of 1.38. In 2002–2005, the fractions of women with PPH showing SI of ≥1.5 with estimated blood loss of 500–1,499, 1,500–2,499, and ≥2,500 ml were 0, 6, and 100%, respectively. In 2013–2015, those fractions were 8, 0, and 17%, respectively.
When we evaluated SI on arrival in women with no transfusion, those with RBC transfusion alone, and those with both RBC and FFP transfusion in 2002–2005 and 2013–2015, SI on arrival was <1.5 in 78% (18/23 cases) in women who received both RBC and FFP transfusion (Table 3). In women with both RBC and FFP transfusion, blood loss on arrival, total blood loss, and maximum PT-INR were significantly greater than in those with no transfusion, and the minimal Hb concentration and minimal fibrinogen concentration were significantly lower than in those with no transfusion.
| Presence/absence of transfusion | Missing data | P-value | Significant pair(s) | |||
|---|---|---|---|---|---|---|
| No transfusion | RBC transfusion alone | Both RBC and FFP transfusion | ||||
| (Group 1, n=36 ) | (Group 2, n=5 ) | (Group 3, n=23) | ||||
| SI on arrival | 0.79±0.20 | 1.00±0.28 | 1.19±0.44 | 0 | <0.001 | 1 vs. 3 |
| SI ≥1.0 on arrival (%) | 4 (11) | 2 (40) | 13 (57) | 0 | <0.001 | 1 vs. 3 |
| SI ≥1.5 on arrival (%) | 0 (0) | 0 (0) | 5 (22) | 0 | 0.008 | 1 vs. 3 |
| Blood loss on arrival (ml) | 1,430±543 | 1,744±566 | 2,166±800 | 0 | <0.001 | 1 vs. 3 |
| Blood loss on arrival ≥1,500 ml (%) | 16 (44) | 3 (60) | 17 (74) | 0 | 0.083 | |
| Blood loss on arrival ≥2,500 ml (%) | 1 (3) | 0 (0) | 7 (30) | 0 | 0.005 | 1 vs. 3 |
| Total blood loss (ml) | 1,762±497 | 2,576±626 | 4,013±1,835 | 0 | <0.001 | 1 vs. 3, 2 vs. 3 |
| Total blood loss ≥1,500 ml (%) | 27 (75) | 5 (100) | 23 (100) | 0 | 0.017 | 1 vs. 3 |
| Total blood loss ≥2,500 ml (%) | 2 (6) | 2 (40) | 22 (96) | 0 | <0.001 | all |
| Minimal Hb concentration (g/dl) | 7.3±1.5 | 5.4±0.5 | 5.4±1.7 | 0 | <0.001 | 1 vs. 2, 1 vs. 3 |
| Minimal Hb <7.0 g/dl (%) | 17 (47) | 5 (100) | 19 (83) | 0 | 0.005 | 1 vs. 3 |
| Maximum PT-INR | 1.04±0.09 | 1.10±0.09 | 1.53±0.72 | 10 | 0.001 | 1 vs. 3 |
| Maximum PT-INR ≥1.25 (%) | 0/32 (0) | 0/4 (0) | 11/18 (61) | 10 | <0.001 | 1 vs. 3 |
| Minimal fibrinogen concentration (mg/dl) | 318±110 | 210±29 | 111±69 | 28 | <0.001 | 1 vs. 3 |
| Fibrinogen on arrival <150 mg/dl (%) | 1/17 (6) | 0/3 (0) | 11/16 (69) | 28 | <0.001 | 1 vs. 3 |
Data are presented as mean±SD, or number (%).
FFP, fresh frozen plasma; PT-INR, international normalized ratio of prothrombin time.
When we evaluated vital signs, SI, and the Hb concentration on arrival at our tertiary center according to the main cause of PPH in 2002–2005, women with uterine inversion showed the poorest conditions: HR and SI were the highest and the Hb concentration was the lowest among the 4 groups (Table 4). The final Hb concentration in those with uterine inversion was the lowest, and all women were transfused. The estimated blood losses were 1,250, 1,410, 1,600, and 4,500 ml in four cases in the group 4; in three quarters of the cases, the blood loss was reported as <2,500 ml; however, SI in all cases was >1.0, and SI was >1.5 in 50%.
| Main cause of PPH | Missing data | P-value | Significant pair(s) | |||||
|---|---|---|---|---|---|---|---|---|
| Atonic bleeding | Placental retention/accreta | Vaginal/cervical lacerations | Uterine inversion | |||||
| (Group 1, n=14) | (Group 2, n=12) | (Group 3, n=6) | (Group 4, n=4) | |||||
| Status on arrival | ||||||||
| Estimated blood loss (ml) | 1,836±354 | 1,505±591 | 1,103±329 | 2,190±1,547 | 0 | 0.045 | None. | |
| HR, (per min) | 98±20 | 87±31 | 106±19 | 142±23 | 0 | 0.005 | 1 vs. 4, 2 vs. 4 | |
| SBP (mmHg) | 122±18 | 109±25 | 127±25 | 88±29 | 0 | 0.038 | None. | |
| SI | 0.81±0.17 | 0.83±0.36 | 0.87±0.24 | 1.71±0.51 | 0 | <0.001 | 1 vs. 4, 2 vs. 4, 3 vs. 4 | |
| SI ≥1.5 (%) | 0 (0) | 1 (8) | 0 (0) | 2 (50) | 0 | 0.012 | 1 vs. 4 | |
| Hb concentration (g/dl) | 8.3±1.7 | 9.4±2.2 | 9.4±2.1 | 5.4±2.8 | 0 | 0.014 | 2 vs.4, 3 vs. 4 | |
| Hb <7.0 (%) | 2 (14) | 2 (17) | 1 (17) | 3 (75) | 0 | 0.064 | ||
| Prognosis | ||||||||
| Total blood loss (ml) | 2,131±465 | 2,774±1,824 | 2,146±609 | 4,131±2,035 | 0 | 0.056 | ||
| The worst value of Hb (g/dl) | 6.7±1.2 | 7.5±1.8 | 5.7±1.1 | 4.3±1.9 | 0 | 0.003 | 2 vs. 4 | |
| Transfusion (%) | 3 (21) | 4 (33) | 2 (33) | 4 (100) | 0 | 0.038 | 1 vs. 4 | |
| Hysterectomy (%) | 0 (0) | 1 (8) | 1 (17) | 1 (25) | 0 | 0.351 | ||
| Maternal death (%) | 0 (0) | 0 (0) | 0 (0) | 0 (0) | 0 | 1.000 | ||
Data are presented as mean±SD, or number (%).
HR, heart rate; SBP, systolic blood pressure.
In most cases, the amounts of infusion were less than twice as much as the estimated blood loss (89% in 2002–2005, and 86% in 2013–2015). We investigated the correlation coefficients among the Hb concentration just after arrival, estimated blood loss, and the amount of infusion before arrival at our tertiary center, using cases in 2002–2005 and 2013–2015. The correlation coefficient between the Hb concentration just after arrival and estimated blood loss was larger than that between the Hb concentration just after arrival and the amount of infusion before arrival at our tertiary center (r: –0.685 vs. –0.369, respectively). We also checked the amounts of fluid resuscitation in two cases showing Hb <7.0 g/dl in the group with estimated blood loss of 500–1,499 ml in 2002–2005. One woman showing a Hb concentration of 5.5 g/dl due to 1,250 ml of blood loss had only an initial infusion of 1,000 ml, and the other woman showing a Hb concentration of 6.5 g/dl due to blood loss of 1,400 ml also received an initial infusion of only 1,300 ml.
We hypothesized that if estimated blood loss is accurately assessed, the Hb concentration on arrival at a tertiary center will be ≥7.0 g/dl in all women with PPH, in whom the estimated blood loss is reported as <1,500 ml. We also hypothesized that if blood loss is accurately estimated, SI on arrival at a tertiary center will be <1.5 in all women with PPH, in whom the estimated blood loss has been reported as <2,500 ml. Our chart review clearly demonstrated that there were some PPH cases showing mismatch between the reported estimated blood loss and the Hb concentration and/or vital signs in 2002–2005, when the guideline for the management of critical obstetrical hemorrhage had yet to be published. Interestingly, obstetrical hemorrhage cases showing mismatch between the reported estimated blood loss and the Hb concentration and/or vital signs still existed in 2013–2015, when the guideline for the management of critical obstetrical hemorrhage appeared to have been sufficiently disseminated among obstetricians.
The Japanese guideline for the management of critical obstetrical hemorrhage was initially published in Apr. 2010;5) the guideline stated that obstetricians should manage obstetrical hemorrhage based on abnormal vital signs (tachycardia, hypotension, and oliguria), especially SI, instead of estimated blood loss, because some diseases related to delivery show relatively low-level external bleeding, but show marked intra-abdominal/retroperitoneal hemorrhage, in which estimated blood loss could not reflect the actual bleeding volumes.5) In 2002–2005 and 2013–2015, mismatch of estimated blood loss and SI in women with estimated blood loss of <2,500 ml occurred in 3.5% (2/56). Therefore, in some PPH cases that showed external bleeding, but not intra-abdominal/retroperitoneal hemorrhage, estimated blood loss might have been underestimated. Thus, our data suggest that the new guideline may be appropriate for all PPH cases with the potential to develop critical obstetrical hemorrhage.
Our data revealed the new problem of whether SI of 1.5 is an appropriate cutoff value for the initiation of RBC transfusion concomitant with FFP transfusion. In a retrospective case-control study involving 50 women with and 50 women without PPH, the mean SI at 30 minutes after the onset of PPH was 0.90 and 0.74, respectively; 75% of women with SI of 1.1 or more at 30 minutes required transfusion, indicating that a cutoff value of SI lower than 1.5 may be appropriate as an indication for RBC/FFP transfusions in women with PPH.16) In another referral center in Japan, 80 emergency referral patients who had received blood transfusion were reviewed; in women with blood transfusion due to genital tract trauma, uterine atony, placental abruption, and uterine inversion, the median SI was 1.043, 0.894, 0.649, and 1.555, respectively.12) Thus, in the majority of women with massive obstetrical hemorrhage, blood transfusion was performed with SI <1.5, suggesting that SI alone may be insufficient for determining blood transfusion. Actually, in our data, SI on arrival was <1.5 in 78% (18/23 cases) of women who received FFP transfusion in 2002–2005 and 2013–2015. Therefore, an appropriate cutoff value of SI in those requiring both RBC and FFP should be reconsidered based on a large, multicenter, cohort study in the future.
Among the 4 main causes of PPH, uterine inversion led to the poorest vital signs on arrival at our tertiary center. In PPH due to uterine inversion, all SI were >1.0, and half of them were >1.5, although three-quarters had an estimated blood loss <2,500 ml. Of course, factors other than bleeding might be significantly correlated with SI for uterine inversion, including parasympathetic nerve reflection. However, transfusion was performed in all cases, and hysterectomy was performed in one case. In another study investigating SI in women with massive critical hemorrhage, SI in uterine inversion was significantly higher than those in genital tract trauma, uterine atony, and placental abruption, whereas the estimated blood loss in uterine inversion was the lowest among the 4 groups.12) Thus, these results mean that uterine inversion cases transferred to a tertiary center tend to show pre-shock or shock even if the estimated blood loss had been reported to be relatively low. Although the incidence rate of uterine inversion is very low, uterine inversion is a life-threatening emergency, because it tends to lead to a shock status due to severe, continuous hemorrhage. For example, in the 40 cases of uterine inversion from 1977 to 2000 in a maternity hospital in Canada, 65% were complicated by PPH, and 47.5% required blood transfusion.17) Our data, although the number of cases was very limited, suggested that we should manage women with uterine inversion very carefully while monitoring vital signs, because the time to progression to critical hemorrhage is very short, and the blood loss is often underestimated.
Theoretically, if estimated blood loss is accurately assessed and appropriate infusion is performed, the Hb level on arrival will be lower; on the contrary, if the amounts of infusion before arrival are not enough to supply an adequate blood volume, the Hb level on arrival will be higher.9) Therefore, it may be dangerous to judge the severity of blood loss or required quantities of blood transfusion based on the Hb level alone. In emergency situations involving pregnant women with obstetrical hemorrhage, which is the better determinant of the Hb concentration on arrival at a tertiary center, estimated blood loss or the amount of infusion before arrival at the center? In our data, the correlation coefficient between the Hb concentration just after arrival and estimated blood loss was larger than that between the Hb concentration just after arrival and the amount of infusion before arrival, suggesting that the estimated blood loss is a more important factor related to the Hb concentration just after arrival than the amount of infusion before arrival at a tertiary center. Actually, there were two women showing mismatch between the estimated blood loss and Hb level in spite of relatively low fluid infusion in the group with an estimated blood loss of 500–1,499 ml. We speculate that the mismatch in the two women may have occurred due to inappropriate calculation of blood loss rather than the blood dilution.
This retrospective study has several limitations. Firstly, the data collected were relatively old. However, we thought that data after publishing the guideline for the management of Critical Obstetrical Hemorrhage are not appropriate for assessing SI for the management of PPH, because the clinicians had already changed the way to decide on the timing of transfusion according to SI from the estimated blood loss after 2010. This retrospective study was a secondary analysis following the first analysis in 2007, when the guideline had yet to be published, to examine whether SI is useful for assessing the estimated blood loss in women with PPH. Secondly, in the survey performed in 2007, we did not collect data on the obstetrical DIC score. Because the medical charts before 2004 had already been discarded, we could not obtain DIC data for this study. Thirdly, it may be better to use the subtraction of Hb concentrations between before and after parturition, rather than the Hb concentration just after arrival. However, in the majority of cases, the values of Hb in the third trimester in the transferred clinic/hospital were not written in the medical referral letter. Therefore, we could not calculate the subtraction of Hb concentrations between before and after parturition.
In conclusion, our chart review clearly demonstrates that there are some PPH cases showing mismatch between the reported estimated blood loss and the Hb concentration and/or vital signs. Therefore, our data support the statement of the Japanese guideline for the management of critical obstetrical hemorrhage published in 2010. Especially, SI may be clinically useful for judging transfusion in women with uterine inversion. We should identify the factors in such mismatching cases, and reveal the pitfalls of the current guideline for critical obstetrical hemorrhage mainly based on SI. However, our data generated the new question of whether the cutoff value of 1.5 for SI is too high for considering RBC/FFP transfusions. Therefore, a large, multicenter, cohort study should be conducted to determine an appropriate cutoff value of SI in those requiring both RBC and FFP in the future.
The authors have no conflict of interest to declare.
